Children with dystonia do not improve their movement performance despite a lifetime of practice. Recent theoretical results suggest that this may be due to one of two types of failure of motor learning: Type 1: inability to recognize relevant errors, Type 2: inability to generate examples of correct behavior. We propose a set of experiments to demonstrate that failure of motor learning may contribute to poor motor control in children with dystonia. To test this hypothesis, 30 children with primary or secondary dystonia and 30 control children will use surface electromyographic activity of either 2 or 4 muscles to control the position of cursors on a computer screen. Type 1 failure can be induced artificially by obscuring visual information about muscle activity. Type 2 failure can be induced using a difficult task in which 4 muscles control 4 dimensions of movement through an unknown linear mixing transformation, and the children must discover exactly one specific pattern of activity. In specific aim 1, we will demonstrate that both control children and children with dystonia show increased co-activation of muscles when information about muscle activity is obscured, as predicted by type 1 failure of motor learning. In specific aim 2, we will demonstrate that both control children and children with dystonia are unable to learn a difficult task until a successful example is learned in a simplified version of the same task. A successful result of this study will show that failure of motor learning is necessary and sufficient to produce part of the motor deficits in childhood dystonia. It will also show strategies for improvement of dystonia. In particular, type 1 failure can be improved ifbio- feedback of an unobserved mode (in this case, co-contraction of biceps and triceps) is provided. Type 2 failure can be improved if the correct solution is presented to children in a simplified task, so that they can then remember and return to the correct solution at will. These experiments will demonstrate a potentially important contributor to motor symptoms in dystonia, and they will indicate specific new treatment opportunities.